Asterisk RTP Loss

Hai Everyone.

I have an asterisk box installed with Asterisk 16.13.0. We have a SIP trunk from the telecom operator configured on the server extra Lan card. The server had 2 land card 1 is configured with private IP and the other one is configured with Telecom operator IP. and also configured route file to map the media gateway of the telecom operator

The issue which is facing the voice issue - Most of my calls are effected by Voice issue- I have taken a pcap trace on the interface where my telecom operator IP configured and during the checks found that Asterisk not sending the RTP to my telecom operator in during the calls. I am doing the conference call using just a local channel and both calls going through my telecom lines.

what could be the reason for the asterisk dropping the RTP packet which was supposed to sent to the telecom lines.

I don’t understand this. If you are between calls, you are not in a call, so there is no RTP.

Asterisk doesn’t drop RTP, on a bridged call, although the underlying OS may drop it if there is no route to the destination of the RTP.

Your trace is showing a very high packet loss, which indicates an overloaded network.

@david551 - Actual let me clear that point first.

Asterisk dropping the RTP packet during the calls- When i enable the pcap trace as per the screenshot almost 17% packet got dropped without reaching the network interface.

Is there any scenario where asterisk drops rtp when there is CPU load or RAM utilization.

Am using the CentOS Linux release 7.6.1810 (Core)

Any dropping should produce a log message.

It will obviously drop packets if the OS refuses to accept them into kernel space. There are cases where the internal frame queues can get very long, but that definitely produces a message, and I’m not sure that it drops the packet.

I wouldn’t expect the frame queue to only drop occasional packets. As such I would say the only sensible reason would be a network overload backing up through the OSes transmit buffers, causing the the system call sending the packet to fault.

That sort of pattern is normally only seen when the network drops the packets after they have been sent.

25% is suspicious. I’m wondering if the OS is sharing the traffic between the two interfaces.

We have a route file which will route the network traffic of my operator IP through specific interface.

This is something which is getting while on the calls- Result of sip show channelstats